Compounds having biological activity can be identified by screening diverse collections (or libraries) of compounds that are produced by organic synthesis, fermentation or molecular biological methodologies. Drug discovery and optimization rely heavily on structure-activity relationships developed by altering the structure of lead compounds and determining the effect of these alterations on the observed biological activity. Complex molecules may require many structural modifications to understand the essential molecular architecture for optimum biological activity and techniques that can be used to accelerate the process are very valuable.
Combinatorial chemistry and compound libraries are valuable tools for both lead discovery and optimization and new methodology to prepare chemical libraries is a continuing need. Combinatorial libraries can be designed to maximize structural diversity or they can be designed to systematically vary substitution around a common chemical core. Because many ligands for biologically important receptors or enzyme active sites are non-peptides there is a continuing need to develop new techniques which will produce greater diversity and useful properties in more conventional small molecule libraries.
A variety of approaches to the preparation of chemical libraries have been developed. Many early efforts focused on using a limited set of high-yielding reactions thereby minimizing by-products and the need for purification steps. Alternatively, multi-component condensation reactions, e.g. the Ugi Reaction, can be exploited to assemble multiple fragments in a single reaction. While these approaches remain useful they fail to exploit many useful reactions which have been developed by synthetic organic chemists.
Syntheses of molecules on a solid support frequently facilitate both the synthesis and purification of chemical libraries. Solid phase synthesis was developed initially for peptide synthesis, and subsequently adapted to oligonucleotides. The range in molecular diversity in these natural biopolymers is relatively limited and the successful implementation of solid phase synthesis required optimization of a limited number of transformations. Small molecule libraries, however, exhibit a vastly greater range a complexity, structural diversity, and chemical reactivity and adaptation of solid-phase synthesis techniques to small molecules requires a significantly larger repertoire of synthetic methodology to efficiently access the diversity inherent in these small molecules. (Thompson, L. A. et al., Chem. Rev. (1996) 96:555; Fruchtel, J. S. et al., Angew. Chem., Int. Ed. Engl., (1996) 35:17; Czarnik, A. W. and Ellman, J. A. (Eds.) Combinatorial Chemistry Special Issue. Acc. Chem. Res., (1996) 29:112)
In solid phase synthesis a first reactant is attached to a solid support. This attachment can be a direct covalent bond to a functional group on the solid support or, alternatively, the attachment can be through molecular spacers or linkers between the solid support and the first reactant. Those spacers can be designed to modify the chemical reactivity of the reactant or to provide routes to ultimately cleave the compound from the solid support. As chemical reactions take place, the intermediate remains linked to the solid support while unreacted reagents or molecules can be removed easily by washing and filtration after each reaction step is completed. This allows large excess of reactants to be used to drive a reaction to a desired product without introducing serious purification problems. Immobilization of the reactant on a solid support produces high-dilution conditions which can promote intramolecular reactions and limit undesired side reactions. These relatively simple manipulations are readily automated which further increases the efficiency of the process. The added cost of the solid support is frequently offset by less labor-intensive purification steps and less need for solvents and adsorbents for chromatography. Multi-step processes can be carried out to efficiently produce complex organic molecules with minimal purification.
Peptides and proteins play a critical role in regulating many biological processes. Unfortunately peptides are susceptible to chemical and enzymatic hydrolysis and are difficult to deliver systemically. This has stimulated the search for biologically active small molecules, peptidomimetics, that mimic endogenous peptides, but are stable to physiologic conditions and are bioavailable after oral administration. Although a variety of scaffolds have been identified which mimic secondary conformations of proteins and polypeptides, the enormous variety of conformations found in nature affords a continuing need to identify useful templates to mimic polypeptides.

The 3,5,5-trisubstituted pyrrolin-4-one ring system, (5a) has proven to be a versatile template for the design of nonpeptide peptidomimetics and polypyrrolinones (21) have been shown to be effective surrogates for polypeptides. Depending on their structure, polypyrrolinones, which are stable to both strong acid and proteases, can adopt diverse conformations including those analogous to β-strands (Smith, A. B., III et al., J. Am. Chem. Soc. 1992, 114, 10672; Smith, A. B., III et al., J. Am. Chem. Soc. 1994, 116, 9947), β-turns and helices (Smith, A. B., III et al., Bioorg. Med. Chem. 1999, 9). The β-strand structural motif was successfully utilized in the design and synthesis of several potent, bioavailable inhibitors of the HIV-1 aspartic acid protease which exhibited improved membrane transport properties relative to their peptidal counterparts. (Smith, A. B., III et al., J. Med. Chem. 1994, 37, 215; Smith, A. B., III, et al., J. Am. Chem. Soc. 1995, 117, 11113; Smith, A. B., III et al., J. Med. Chem. 1997, 40, 2440; Thompson, W. J., et al., J. Med. Chem. 1992, 35, 1685.) The improved transport was attributed to the presence of intramolecular hydrogen bonds between adjacent pyrrolinone rings (NH and CO), which led to a reduction in desolvation energy upon membrane transport (Hirschmann, R., et al. In New Perspectives in Drug Design; Dean, P. M., Jolles, G., Newton, C. G., Eds.; Academic: London, 1995; pp 1-14.). A bis-pyrrolinone was successfully used in the construction of a pyrrolinone-peptide hybrid ligand, which bound the Class II MHC protein HLA-DR1 in an extended β-strand-like conformation with similar potency to the native peptide. (Smith, A. B., III, et al. J. Am. Chem. Soc. 1998, 120, 12704; Smith, A. B., III; et al. J. Am. Chem. Soc. 1999, 121, 9286.)
Recent observations suggest that the polypyrrolinone structural motif, designed initially to mimic peptide and protein β-strand/β-sheet structural motifs may in fact represent a privileged nonpeptide scaffold, able to mimic not only the extended β-strand/β-sheet conformation, but also other diverse conformations including those analogous to β-turn and helices. This unexpected diversity, if accessible in controlled fashion, would expand the scope of the polypyrrolinone scaffold for the development of low-molecular weight ligands for a variety of biologically important targets.
An iterative solution phase syntheses of polypyrrolinones has been developed (FIG. 1) based upon the intramolecular cyclization of a metalloenamine derived from an α-amino acid derivative. Condensation of a latent 4-oxo-2-aminobutyrate derivative (1) with an aldehyde (2) produces the key imine (3). Deprotonation of (3) with KHMDS and subsequent intramolecular cyclization upon stirring the resulting potassium salt (4) at room temperature produces the 3,5,5-trisubstituted pyrrolinone (5). The olefinic (5a) or acetal (5b) side chains can be oxidatively or hydrolytically transformed to a new aldehyde (Sc) which can be subjected to further iterations of the same reaction sequence to produce polypyrrolinones (6) (Smith, A. B., III, et al., J. Am. Chem. Soc. 1992, 114, 10672; J. Am. Chem. Soc. 1994, 116, 9947; J. Am. Chem. Soc. 1999, 121, 9286; U.S. Pat. No. 5,514,814; U.S. Pat. No. 5,770,732; U.S. Pat. No. 6,034,247; all incorporated herein by reference in there entirety). While this synthetic scheme is adequate to prepare limited numbers of analogs using conventional solution phase techniques, it is not well suited for the rapid preparation of large numbers of polypyrrolinones required for a pyrrolinone library. An iterative, solid-phase synthetic strategy is ideal to automate this multistep sequence.
The existing methodology (FIG. 1) required either a two-step oxidation [(a) OsO4/NMO (b) NaIO4; Approach A] or strong acid hydrolysis [Approach B] to unmask the aldehyde moiety at the outset of each successive synthetic cycle. It has now been found that the solid support is incompatible with a repetitive osmium catalyzed hydroxylation/oxidation process. While a single treatment was successful, subsequent hydroxylations failed. Moreover, ozonolysis, an alternative oxidative procedure is poorly adapted for automated synthesis. Acid hydrolysis of the acetal required increasingly strenuous conditions during each iteration of the cyclization process resulting in unacceptable degradation in yield and product purity. Thus it is apparent that further improvement in existing methodology is need to produce chemical libraries of these valuable compounds utilizing solid phase synthesis techniques.